This paper evaluates the use of energy recovery devices in sea water reverse osmosis. A Pressure exchanger is theoretically compared to other energy recovery configurations resulting in an energy recovery of around 60%. A test rig was implemented by first creating a CAD model of a pressure exchanger on solid edge, manufacturing it and implementing it in a small scale reverse osmosis system. Future work includes fixing minor defects with the test rig in order to carry out experimentation.
Key Words: Reverse Osmosis, Energy Recovery, Pressure Exchanger
Fresh water is defined as containing less than 1000 mg/L of salts or total dissolved solids (TDS). Above 1000 mg/L, properties such as taste, color, corrosion propensity, and odor can be adversely affected. [1]
With reference to fig. 1, The U.S. Geological Survey found that 97% of Earth’s water is located in the ocean, the remaining 3% makes up for the fresh water composition. Out of this 3%, around 2% of it is located in the ice caps and glaciers; the remaining 1%t includes surface water which comprises of swamps, lake and rakes. The remaining percentage is made up of brackish water, slightly salty water found as surface water in estuaries and as groundwater in salty aquifers. [1]
Today, the production of potable water has become a global challenge. With reference to fig. 2 projected population growth and demand exceed conventional available water resources. At present, around 1 billion people are without access to clean drinking water and approximately 40% of the world population lives in water shortage regions. [2]
Increasing demand & decreasing supply of water has led to ideas such as water conservation and water transfer or dam construction being implemented although they are still not sufficient to cope up with the population growth. Misuse or overuse of traditional fresh water resources such as lakes, rivers, and groundwater result in them either diminishing or becoming saline. At present due to global development, the introduction of few new water resources are available to support daily clean drinking water needs. The facts indicated above clearly show us that salt water desalination has emerged as the prime candidate to provide fresh drinking water to sustain future generations across the globe. According to a recent graph by ERI Inc., by 2016 capital expenditures for desalination will exceed 16 billion$ out of which more than 13$ is expected to be targeted for SWRO. This massive injection of finances will successfully incorporate additional clean drinking water production for all sorts of communities using conventional water treatment and fresh water resources. [3], [4], [5], [6], [7], [8]
A solvent moves from an area of low solute concentration, through a membrane, to an area of high solute concentration through osmosis. The movement of a pure solvent to equalize solute concentrations on each side of a membrane generates a pressure and this is the "osmotic pressure." With reference to fig 3. applying an external pressure to reverse the natural flow of pure solvent, thus, is reverse osmosis. [10]
Fig. 4 represents a typical reverse osmosis system which comprises of a high pressure pump and a membrane. The High pressure pump pressurizes sea water through a reverse osmosis membrane at pressure of approximately 60 bar as this is what’s needed in order to overcome the osmotic pressure. 40% of the flow from the membrane results as fresh water output while the remaining 60% is concentrated brine that is rejected with a great pressure loss. This pressure loss is fed back to the system using energy recovery devices which are detailed in 1.4.
The energy consumption still remains the major operational cost component due to the high pressure pumps required to feed water to the RO process. These pumps are responsible for more than 40% of the total energy costs [2]. Reducing energy consumption is, therefore, critical for lowering the cost of desalination and addressing environmental concerns. [9]
Before the concentrate stream is sent for disposal, pressure from the stream is recovered by passing it through an ERD. When an ERD is used, a fraction of the feed must bypass the primary high-pressure pump and a booster pump is used to account for pressure losses in the RO membrane modules, piping, and ERD. The fraction of power recovered depends on the type and efficiency of the equipment used. [11]
Greenlee et al. (2009) & Wang et al. (2004) discussed the two broad classes of ERDs. Class I devices use the hydraulic energy of the RO concentrate by first converting the energy to centrifugal mechanical energy and then back to hydraulic energy. This is a two-step process. Class II devices use hydraulic power to cause a positive displacement within the energy recovery device and this transfers the hydraulic energy directly in a one step process. [12]
Stover (2007) provided a detailed description of Class I Centrifugal ERDs (such as the pelton wheel, reverse running turbine pump and turbo charger). They are limited in capacity and are usually optimized for narrow flow and pressure operating conditions. Initially, Francis turbines were applied, but they were replaced by pelton turbines that operated at higher efficiency in high-head applications and provided a maximum possible efficiency of 90% [13] Oklejas (2005) mentioned the turbo charger consists of an impeller and turbine on the same shaft, this is typically used in smaller capacity RO installations as its efficiency ranges from 55% to 60%. [14].
Mirza (2008) discussed mechanically coupled reverse running turbine pump that have efficiencies in the range of 75% to 85%. For the submersible generator type, the overall efficiency is in the 62% to 75% range. Therefore, this type of ERD is not suitable for a low flow range. [16]
At present, most of the desalination plants use a Class II type of ERD namely a pressure or work exchanger that can achieve efficiencies greater than 95% (Greenlee et al. 2009). The PWE transfers the hydraulic energy of the pressurized RO concentrate stream to the RO feed water stream. [13],[15] PWE systems can be categorized as two types: those that provide a physical barrier (piston) between the RO concentrate stream and feed side of the system, such as a Dual Work Exchanger Energy Recovery (DWEER), and those without a physical barrier such as a Pressure Exchanger (Cameron and Clemente 2008; Mirza 2008). In the case of a DWEER, the system is based on moving pistons in cylinders which is well suited for a wide range of water viscosities and densities, but results in a large foot print (Mirza 2008). A Pressure exchanger device has higher efficiency since no transformational losses occur in the device, higher capacity is achieved by arranging several devices in series. Disadvantages include limited flow rates, high noise levels requiring a sound abatement enclosure (Mirza, 2008) and the degree of mixing that occurs between the feed water and concentrate stream. A feed salinity increase of 1.5%e3.0% caused by such mixing will increase the required feed pressure for the RO system (Wang et al. 2005). [16],[17]
M. Barreto et. al. (2010) worked on a RO kinetic energy recovery system which is in the form of a closed loop. feed water accumulated in one of the pressure exchangers is pressurized with a high pressure pump (HPP) by making a closed circuit between the membrane output and the reverse osmosis module input, where a water tank (pressure exchanger) and a booster pump are fitted into the line. It also consists of inertia valves, expansion bladder. The water entering or leaving them, which must be pressurized or depressurized, is always in continual motion to avoid unnecessary consumption of kinetic energy that arise from stops in the operation . Certain advantages include: Reduced specific energy consumption & total cost, 97% energy efficiency, reduced capacity of the HPP, Decrease in the amount of antiscalant needed and low mixing percentage between seawater and brine in the isobaric chamber. [5]
Xiaopeng Wang et. Al. (2010) worked on a positive displacement (PD) ERD known as the FS-ERD that was mainly composed of three portions, a rotary fluid switcher, two pressure cylinders and a check valve nest. The rotary fluid switcher was the core component that consisted of four joint ports and two working phases similar to a two position four-way valve. When the FS-ERD accomplishes its pressurizing stroke (and also the depressurizing stroke), the switcher would rotate to working phase II at a low speed of 7.5 rpm driven by motor, which denotes that the stroke modes in cylinders are alternated to each other. The switcher accomplishes its phase change by rotating its multi-channel rotor around the switcher's shell. Single alone and parallel operation (Flow rate and pressure fluctuations) tests showed that the parallel operation of two sets of ERDs can not only extend the capacity of the system but also remarkably improve the stability and continuity of the working streams to and from the ERDs. The maximum recovery efficiency achieved is 95%. The long contact time (20 to 60 seconds) between the brine and seawater in the isobaric chambers results in some intermixing, resulting and an increase in the membrane feed salinity of up to 1.5%. [18]
From the literature review it is clear that isobaric ERDs deliver higher efficiency than centrifugal devices, but centrifugal devices are generally better characterized and are easier to maintain and operate. Rotary isobaric devices provide a unique combination of isobaric and centrifugal features with high energy transfer efficiency, no maintenance, and easy operation
The pressure exchanger comprises of a Rotor (the only moving device) that rotates about a longitudinal axis and has a plurality of continuous rotor channels having openings on each rotor end face arranged around the longitudinal axis of the rotor with the rotor channels communicating with the connection openings of the housing via flow openings formed in the housing such that during the rotation of the rotor the rotor channels alternately carry high pressure liquid and low pressure liquid from the respective first and second liquid systems. This is enclosed in housing along with hydrodynamic bearings. On either side of the rotor, end covers are fixed with inlet and outlet connection openings for each liquid. The end covers and rotor are enclosed in a sleeve. [13]
With reference to Figure 3, Low pressure sea water enters the pressure exchanger and fills rotor, this sea water is then exposed to high pressure concentrate from the membrane. Pressure transfers directly from the concentrate to the sea water inside the rotor ducts. Spent concentrate leaves the rotor ducts as it gets pushed out by low pressure sea water. The ducts of the rotor functions like a carousel charging and discharging. Water around the narrow gap in the rotor serves as a lubricant. [13]
In the design stage for inlets into the rotor channels, the flow ratios are based on velocity triangle diagrams in which the circumferential component c u generates a driving torque for the rotor as a momentum force. This circumferential component is designed to be larger than the circumferential velocity U of the rotor. The rotor inlet edges formed between the openings of the rotor channels with the wall surfaces which follow in the direction of flow are constructed so that the resulting relative flow of the rotor is received without impact by the rotor channels and is deflected in the direction of the rotor channel length.
Such a design of the inlet of the rotor channels also includes the advantage that when there is a change in volume flow, the triangle diagram of the velocity at the inlet of the rotor channels undergoes an affine change, i.e., the circumferential component c u changes to the same extent as the oncoming flow velocity c of the liquid. Thus the driving torque acting on the rotor also increases, leading to an increase in the rotor rpm. With an increase in rotor rpm, the frictional moment acting on the rotor and having a retarding effect also increases. Due to the linear relationship between the driving torque M I which increases with an increase in the circumferential component c u and the frictional moment M R which increases in proportion to the rotational speed, the circumferential velocity of the rotor is always established so that the triangle diagrams of the velocity conditions which prevail at the rotor inlet are similar for all volume flows. There is thus a self-regulating effect which guarantees the condition of impact-free oncoming flow for each volume flow established. The rotational speed of the rotor is thus corrected based on the congruent velocity triangle diagrams and an impact-free oncoming flow of the rotor channels for volume flows of the main flows that are altered due to system conditions.
A rotor is constructed in multiple parts, whereby a rotor part having straight rotor channels on its end faces is provided with one or two incoming flow plates, and inlet openings and/or downstream channel beginnings which make the channel flows uniform are arranged in the incoming flow plates.
Rotor channels having a trapezoidal cross section are arranged so they are axially parallel to and concentric with the axis of rotation of the rotor , with wall surfaces designed as webs running radially between the rotor channels extending between the rotor channels. The openings in the rotor channels arranged on the end face of the rotor have additional rounded surfaces on their radially outer corners in the manner of inclined surfaces that widen diagonally outward, so that each opening is slightly enlarged.
Opposite the openings of the rotor with its axially parallel rotor channels, The velocity triangle diagram for a liquid flowing into the rotor, comprising velocity vectors U, w and c, where the arrows indicate the directions and the magnitudes of the various velocities, where:
U=circumferential velocity of the rotor
w=relative flow in the opening upstream from the rotor channel
c=absolute flow of the liquid flowing out of the housing and to the rotor, where:
c u =circumferential component of the absolute flow and
c x =axial component of the absolute flow,
?c u =driving velocity for the rotor=c u -U
a=angle of flow of the absolute flow c
ß=angle of flow of the relative flow
The flow to the rotor 1 is passed through a housing part opposite the rotor (not shown) which is opposite the rotor so that the flow in the stationary reference system strikes the rotor 1 as an absolute flow c at the angle a. The rotor 1 rotates with the circumferential velocity U and accordingly the relative flow w strikes it at the angle ß. The circumferential component c u of the absolute flow c is greater by ?c u than the circumferential velocity U of the rotor, thus ensuring the required driving torque of the rotor.
Because of the relative oncoming flow angle ß, which is different from zero, the oncoming flow of the rotor channels in the relative system is not free of impact. Consequently, separations in the form of eddies are constantly developing in the openings in the rotor channels and as a result an irregular velocity profile is established within the flow in the remaining path of the rotor channels. These irregular velocity profiles lead to the mixing problems.
(Kochanowski, 2007)
Reverse osmosis systems consisting of a pressure exchanger, pelton turbine and no energy recovery device have been analyzed below. Block diagrams of each energy recovery device had been constructed to asses and analyze the forces acting on the fluid.
The fluid flow through the rotor channels is viscous.
Mixing occurs in the rotor channels
Eg. Efficiency =
Mixing =
A
B
C
D
E
F
G
H
m3/day
130
111
19
111
130
13
117
117
bar
2.5
2.5
62.0
59.8
62.0
0.0
60.4
2.0
ppm
35,000
35,000
35,000
35,436
35,616
200
39,551
39,137
Fresh Water Output
13 m3/day
Membrane recovery rate
10%
Membrane feed pressure
62.0 bar
Membrane differential pressure
1.6 bar
Pressure Exchanger Low Pressure discharge pressure
2.0 bar
Feedwater salinity
35,000 mg/l
Motor frequency
50Hz
Cost of power
0.10 $/Kwh
High pressure Pump efficiency
90%
High pressure Pump motor efficiency
87%
Booster pump efficiency
48%
Booster pump motor efficiency
88%
Booster pump VFD efficiency
97%
Pump efficiency
90%
Motor efficiency
87%
Power consumed
1.7 KW
Pump efficiency
48%
Motor efficiency
88%
VFD efficiency
97%
Power consumed
0.7KW
Unit flow
4.9 m3/hr
Lubrication per array
0.2 m3/hr
Lubrication flow
5%
Differential pressure High Pressure side
0.6 bar
Differential pressure Low Pressure side
0.5 bar
Efficiency
93.7%
Mixing at membrane feed
1.8%
Operating capacity
71.6%
Power Savings
9.0KW
Estimated CO2 Savings
47tons/year
Specific power consumption
kWh/m3
4.34
Power cost saved with the pressure exchanger
$/year
7,913
Recovery
10%
Membrane differential
bar
1.6
A
C
F
G
m3/day
130
130
13
117
bar
2.5
60.9
0.0
59.3
mg/l
35,000
35,000
200
39,551
Mechanical energy recovered
0.0 KW
HP pump shaft power
15.4 KW
Motor shaft power
15.4 KW
Motor electrical power
17.2 KW
HP Pump efficiency
57%
HP Pump motor efficiency
90%
Net transfer efficiency
57%
Total power consumption
17 KW
Specific power consumption
31.75 KWh/m3
A
E
F
G
H
m3/day
130
130
13
117
117
bar
2.5
60.9
0.0
59.3
0.0
mg/l
35,000
35,000
200
39,551
39,551
Turbine efficiency
56%
HP Pump efficiency
57%
HP Pump motor efficiency
90%
Net transfer efficiency
32%
Mechanical energy recovered
4.5KW
HP pump shaft power
15.4KW
Motor shaft power
10.9KW
Motor electrical power
12.2KW
Recovery
10%
Membrane differential
1.6 bar
Total power consumption
12 KW
Specific power consumption
22.47 KWh/m3
Power saved with Pressure Exchanger
18.13 KWh/m3
81%
Cost saved with Pressure Exchanger
8,604
Based on the tabulated values give above a certain set of trends and variations can be observed
Considering all the aspects taken into account for the direction of this project it seems the results obtained are viable.
With reference to fig 22. as the recovery rate increases the brine concentration, and the membrane flow is insufficient to remove the salts that deposit on the membrane surface. This in turn increases the pressure drop, thus increasing the HPP energy consumption by decreasing the efficiency of the pressure exchanger.
With reference to fig. 23, the pressure drops and viscous friction associated with the pressure exchanger can be explained
The pressure of the feed water flowing from the Pressure exchanger is slightly lower than the pressure of the brine fed to it.
Similarly, the pressure at the brine outlet of the Pressure exchanger is slightly lower than the pressure at the feed water inlet.
With reference to fig. 24, as the flow rate increases, the mixing that takes place between the high pressure sea water exit of the pressure exchanger and the sea water pressurized by the high pressure pump reduces.
Mixing is one of the biggest issues in a pressure exchanger design even though trapezoidal channels are employed.
With reference to fig. 25, the rotor sits on hydrodynamic bearings. Around the rotor a narrow gap filled with water serves as lubrication that helps it spin at a constant rate of approx... 1200 rpm. At a higher flow rate the amount of lubrication provided to the hydrodynamic bearings reduces and settles to around 3.5% at around 200 m3/day
With reference to fig. 26, fig. 27, fig. 28 and fig. 29 the amount of energy recovered by the pressure exchanger is almost 30% more than what a conventional pelton turbine can achieve followed by a saving of almost 60% in a system with no energy recovery device. The high pressure pump accounts for almost 70% of the energy in the reverse osmosis process, introducing a pressure exchanger reduces the energy consumption as compared to any other system. Lower flux rates and lower recovery rates generally result in lower system energy consumption. Fitting a pressure exchanger in plants without any energy recovery would result in massive savings/year. This would even encourage new businesses and would eventually lead to water abundance!
As the recovery ratio increases, the power cost saved decreases; an optimal system would function at 40% recovery.
With reference to fig. 30, The limit is 80%; past that a trend is observed where in the pressure exchanger efficiency is lesser than that of the pelton wheel for higher recovery ratios.
The two most important measures of energy recovery device performance are energy transfer efficiency and concentrate-feedwater mixing, both of these have been met at a very high scale using the pressure exchanger energy recovery device.
Compared to older energy recovery systems, RO systems consume 15 to 35% less power with pressure exchangers
Reduced High pressure pump consumption and system power consumption drop.
A restricted operating range and mixing of the two liquids found in the rotor channels during operation.
7 Conclusions
The global water crisis has reached such a stage where action is needed right now. This paper looks into recovering energy for a typical reverse osmosis in order to make it more affordable and efficient. From the results presented above it is quiet clear that the pressure exchanger fulfils what has been mentioned.
From the literature review, we gather key information about the different types of energy configurations used. It was necessary to understand how a reverse osmosis system works. Each part in detail. Assume certain parameters and calculate system outputs.
A block diagram of the reverse osmosis system with 3 different energy recovery configurations was drawn. Certain parameters regarding efficiency, feed pressure...etc were assumed. Next these inputs values were fed into the system, a set of equations were used in order to calculate flow, pressure...etc at each point and ultimately the system power consumption and performance. The analysis shows that the pressure exchanger recovery system recovers almost 95% of the energy wasted in the brine.
To further investigate this proposal, a CAD model of the pressure exchanger was built in CAD, drawings were obtained and it was manufactured using acrylic. A household reverse osmosis unit was purchased for testing. Obtaining experimental results from the test rig would be ideal to validate the theoretical results. A real life model would include losses as well which would provide useful insight.
A CAD model of the pressure exchanger was created in solid edge; drawings for each part were produced and handed over to the workshop where manufacturing took place using acrylic.
A rod was purchased to be inserted through the apparatus. On either end covers pipe fittings for the LP, HP Sides were connected.
Silicon Glue was used to place the parts in place.
The system would be tested out by measuring the flow, pressure and salinity at each point. This would be later validated by the theoretical results obtained.
A reverse osmosis system consisting of a pressure exchanger was modeled as shown in figure 5, 6 and 7.
Perform experimentation and obtain values for pressure, flow rate and salinity at each point.
Install fluid bearings (hydrodynamic) to the pressure exchanger model.
Replace the existing pipes and fittings with larger diameter counterparts.
First of all I would like to thank my supervisor Dr. Sarim for not only being a great supervisor but for being a great friend. Mr. Mohamed the lab technician also deserves a special note as it was with him that the test rig was able to be set up. Credit goes out to my colleagues for guiding me in case of any hurdle faced.
This is Preview only. If you need the solution of this assignment, please send us email with the complete assignment title: ProfessorKamranA@gmail.com
